Insight into the power of the new Additive Manufacturing File (AMF) format
In a recent post on The End of Polygons, Aaron Trocola correctly points out that the STL file format is becoming a major barrier to production of objects with high volumetric complexity, such as internal lattices and graded materials – essentially reaching a point where STL files make certain advanced applications difficult or impossible.
Recognizing this limitation, several manufacturers and CAD vendors have recently agreed on a new XML-based open standard for additive manufacturing called AMF (Additive Manufacturing File Format). Like STL, the AMF allows descriptions of objects using a mesh of triangles. But there the similarity ends. AMF allows for describing many more details, such as the color of the surface, as well as the volumetric structure of the interior. Skipping color specification for a moment, let’s focus on volume specifications.
As Trocola points out, different representations work well for different situations. Voxel based approaches are suitable for some applications such as for models derived from medical imaging, but in general, they are challenged in both their scalability and their resolution dependency. Functional representations work well for structures described by equations, but they are difficult to implement for arbitrary geometries where mesh representations shine.
AMF attempts to solve this dichotomy by offering a hybrid representation. If you’re only interested in a solid object you can stick to a triangular mesh (see simple AMF). But if you want a lattice structure or graded materials, there are a number of ways to specify them using a combination of voxels and function representations.
In AMF, the internal structure inside a mesh surface is described by defining meta-materials. A primary material is defined as one of the base solid materials that can be produced by a 3D printer. A meta-material is defined by combining primary materials (as well as voids) in various combinations. Materials are combined using mixing coefficients: A constant mixing coefficient makes a simple interpolation of the primary materials. But a coordinate-dependent mixing coefficient can make a graded material – linearly or nonlinearly graded. A periodic mixing coefficient can make lattices, and from there, the sky is the limit. This is both compact and resolution independent, much like the immersed boundary element that Trocola mentions. See some examples here.
AMF also allows meta materials to combine the primary material by referencing a 3D-voxel map, much like texture mapping is used to paint bitmaps onto curved surfaces. Combining functional representations, voxels representations, and mesh representations allows both CAD software and 3D printers quite a bit of freedom.
When will AMF be adopted?
AMF in its recently-approved second revision awaits adoption by major equipment manufacturers and CAD vendors, and/or by the maker community. The specification is described on Wikipedia, and an open-source, license-free implementation of most of the features is available on the AMF wiki.
Yet, like the chicken and the egg, CAD vendor wait to see if equipment manufacturers will support the new format, and equipment manufacturers await a “Save AS” option in major CAD software.
The entire additive manufacturing field could leap forward once both CAD vendors and equipment manufactures adopt the new format. The companies that have the most to gain are those equipment manufacturers whose advanced 3D printers’ capabilities are not used to their full potential because of STL limitations (such as color printing, multi-material printing and lattices). Also standing to gain are new CAD systems that can create advanced geometries but not efficiently export them in STL.